Weaker lewis acid, better catalytic activity: Dual mechanisms in perfluoroarylborane-catalyzed allylstannation reactions

Article abstract of DOI:10.1021/ol034928i

(Matrix is presented) PhB(C₆F₅)₂ exhibits much higher activity as a Lewis acid catalyst for the allylstannation of aromatic aldehydes than the stronger Lewis acid B(C₆F ₅)₃. This anomalous enhancement of catalytic activity for the weaker LA is shown to be partly due to decreased thermodynamic stability of ion pair 2b relative to 2a in the product-forming step of the reaction. A mechanistic path where the borane serves as the true LA catalyst is more important for the weakly Lewis acidic borane.

Full text of DOI:10.1021/ol034928i

ORGANIC
LETTERS
2003
Vol. 5, No. 16
2857-2860
Weaker Lewis Acid, Better Catalytic
Activity: Dual Mechanisms in
Perfluoroarylborane-Catalyzed
Allylstannation Reactions
Darryl J. Morrison and Warren E. Piers*
Department of Chemistry, UniVersity of Calgary, 2500 UniVersity DriVe N.W.,
Calgary, Alberta, Canada T2N 1N4
wpiers@ucalgary.ca
Received May 27, 2003
ABSTRACT
PhB(C F5)2 exhibits much higher activity as a Lewis acid catalyst for the allylstannation of aromatic aldehydes than the stronger Lewis acid
6
B(C F5)3. This anomalous enhancement of catalytic activity for the weaker LA is shown to be partly due to decreased thermodynamic stability
6
of ion pair 2b relative to 2a in the product-forming step of the reaction. A mechanistic path where the borane serves as the true LA catalyst
is more important for the weakly Lewis acidic borane.
Lewis acids (LAs) enhance the rates of reactions involving
carbonyl functions via coordination of an oxygen lone pair,
which activates the CdO group toward attack by weak
nucleophiles.1 One of the precepts of LA catalysis is that,
for isosteric LAs, greater LA strength translates into more
effective activation of the CdO group and therefore higher
rate enhancement.2 Although this is likely the case on a
fundamental level, complex mechanisms of operation can
make this false on a practical level.3
Recently we reported the results of a detailed spectroscopic
investigation on the mechanism of B(C6F5)3-catalyzed al-
lylstannation of ortho donor substituted aromatic aldehydes.4
The picture that emerged showed that the majority of
catalysis in this system is carried out by the stannylium cation
in the ion pair 2a depicted in Scheme 1. The borane is largely
sequestered in the anion of this species once the reaction is
initiated by allylstannation of substrate-borane adduct 1a.
Thus, in the presence of excess allylstannane, as would be
(2) Lewis acid strength has been shown to directly correlate with
reactivity in LA-catalyzed reactions of carbonyls. (a) Ene reactions: Laszlo,
P.; Teston-Henry, M. Tetrahedron Lett. 1991, 32, 3837. Laszlo, P.; Teston,
M. J. Am. Chem. Soc. 1990, 112, 8750. (b) Aldol reactions: Kobayashi, S.
Eur. J. Org. Chem. 1999, 15 and references therein. (c) Classification of
LAs on the basis of activity and selectivity: Kobayashi, S.; Tsuyoshi, B.;
Nagayama, S. Chem. Eur. J. 2000, 6, 3491.
(3) (a) La(OTf)3 and Sm(OTf)3 were more effective than the stronger
LA Sc(OTf)3 (see ref 3b) in LA-catalyzed Baylis-Hillman reactions:
Aggarwal, V. K.; Mereu, A.; Tarver, G. J.; McCague, R. J. Org. Chem.
1998, 63, 7183. (b) Fukuzumi, S.; Ohkubo, K. J. Am. Chem. Soc. 2002,
124, 10270.
(1) (a) Lewis Acids in Organic Synthesis; Yamamoto, H., Ed.; Wiley-
VCH: Weinheim, 2000. (b) Shambayati, S.; Crowe, W. E.; Schrieber, S.
L. Angew. Chem., Int. Ed. 1990, 29, 1053.
(4) Blackwell, J. M.; Piers, W. E.; McDonald, R. J. Am. Chem. Soc.
2002, 124, 1295.
10.1021/ol034928i CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/15/2003

Scheme 1
Scheme 2
giving a value of 18.0(7) eu for ∆S and 7.21(6) kcal mol-1
for ∆H.
[1][4][ArCHO]
Keq
)
(1)
[2]
Although the conditions of this experiment do not mimic
those of the catalytic reaction directly, this equilibrium
represents the product-forming step in a borane-catalyzed
mechanism and offers a rare opportunity to assess the
thermodynamic factors affecting this process. The thermo-
dynamic data suggest that the product-forming step in path
b is largely entropically driven and is enthalpically disfavored
perhaps as a result of the fact that the high Lewis acidity of
B(C6F5)3 stabilizes the ion pair 2a via a strong B-O bond.
A weaker Lewis acid might be expected to make path b more
enthalpically favorable. Indeed, the partially fluorinated LA
PhB(C6F5)2,5 which is a weaker LA than its fully fluorinated
the case under catalytic conditions, 2a undergoes allylation
at -40 °C via aldehyde activation by the Lewis acidic
stannylium cation giving ion pair 3 (path a, Scheme 1).
o-Anisaldehyde substrate regenerates 2a by displacement of
stannyl ether product 4 and trapping of the monoligated
intermediate 5.
The traditionally expected product-forming path based on
exclusive borane catalysis involves direct collapse of ion pair
2a (path b). Previously, we observed that in the absence of
excess allylSnBu3 product formation from 2a is slow at -40°
but begins to occur upon warming from this temperature.
Interestingly, in the presence of excess aldehyde substrate,
formation of 4 via path b is significantly suppressed. Closer
examination of this system reveals that this product-forming
step is reversible and this equilibrium can be studied when
allyl stannane reagent is absent.
Ion pair 2a can be generated via the treatment of 1a with
1 equiv of allylSnBu3 in the presence of 2 equiv of
o-anisaldehyde as shown in Scheme 2 (R ) C6F5). At
temperatures above -40 °C, onset of the “path b” equilibrium
is observed; Keq can also be approached from the right side
of the equation by mixing stannyl ether 4 with o-anisaldehyde
and 1a in the appropriate ratio.
The equilibrium constant (eq 1) can be evaluated at various
1
temperatures by using H and 19F NMR spectroscopy to
Figure 1. van’t Hoff plots of -ln(Keq) versus 1000 T-1 for the
equilibria in Scheme 2: (O) B(C6F5)3, (∆) PhB(C6F5)2.
determine the concentrations of the components. A van’t Hoff
plot for this equilibrium is shown in Figure 1 (open circles),
2858
Org. Lett., Vol. 5, No. 16, 2003

The higher activity observed for the weaker LA extends
to other aromatic aldehyde substrates (Table 1). Benzalde-
hyde itself is a more active substrate, but at both -44 and
-78 °C PhB(C6F5)2 is a more effective catalyst than
B(C6F5)3.8 The same trend exists for p-chlorobenzaldehyde,
while the difference is less pronounced for the highly
activated substrate p-nitrobenzaldehyde at -78 °C. This latter
observation is likely a reflection of the effect of the para
substituent on the thermodynamic stability of the ion pairs
2 and suggests that the balance between path a or b is also
affected by the nature of the substrate.
Thus, although B(C6F5)3 is clearly a stronger LA than
PhB(C6F5)2,6 this leads not to higher activity as an allylation
catalyst but rather to greater thermodynamic stability of
the product of the first C-C bond forming step in the
reaction, 2a. The ion pair 2b, on the other hand, does not
fall into as deep a well and the equilibrium is more biased
toward product 4 and 1b, which in the presence of excess
allyltin reagent is rapidly allylated, shunting the reaction
toward path b. However, while the data for the equilibria of
Scheme 2 provide a partial explanation for the observed
higher activity for the weaker Lewis acid PhB(C6F5)2,
clearly there are ill-defined kinetic factors contributing to
the higher rate of transfer of the alkoxy group from boron
to tin in 2b as opposed to 2a. Possibly, the substitution
of one C6F5 group for a less electron-withdrawing C6H5
renders the alkoxy group significantly more nucleophilic
in the phenyl-substituted borate anion, lowering the bar-
rier to product-forming transfer to tin. Given that the in-
timate mechanisms of these equilibria likely involve mul-
tiple steps, the factors affecting the observed rates are
complex.
In summary, unlike B(C6F5)3, the weaker LA PhB(C6F5)2
likely functions as a true LA catalyst for the allylstanna-
tion reaction and is in practical terms much more effective
than either B(C6F5)3 or [Bu3Sn(L)]+. In effect, B(C6F5)3
is partially buffered to the strength of [Bu3Sn(L)]+, and
as previously determined,4 a significant portion of the
catalysis occurs via path a for this borane. PhB(C6F5)2 is
strong enough to promote allylation but weak enough to more
effectively transfer the OR group to tin to complete the
reaction via path b. This concept can only be extended so
far, however; the substantially weaker LA BPh3 (Child’s
Lewis acidity ) 0.06) is not an active catalyst for the
allylstannation of benzaldehyde at or below -44 °C, so a
threshold borane Lewis acidity is required in order to initiate
this reaction.9
Table 1. Lewis Acid Catalyzed Allylstannation of
Benzaldehydea
entry
catalyst
B(C6F5)3
PhB(C6F5)2 o-MeO-C6H4 -44
Ar
T (°C) time (min) convn (%)b
1
2
o-MeO-C6H4 -44
75
75
21
100
95
100
62
96
38
100
95
100
3
4
5
6
7
8
9
B(C6F5)3
PhB(C6F5)2 C6H5
B(C6F5)3 C6H5
PhB(C6F5)2 C6H5
C6H5
-44
-44
-78
-78
-78
-78
60
40
600
600
300
300
70
B(C6F5)3
p-Cl-C6H4
PhB(C6F5)2 p-Cl-C6H4
B(C6F5)3
p-NO2-C6H4 -78
10
PhB(C6F5)2 p-NO2-C6H4 -78
70
a Conditions: CH2Cl2, 0.1 M in ArCHO, 0.5 mmol ArCHO, 0.55 mmol
allylSnBu3, 5.0 mol % catalyst loading. b Conversion of ArCHO by GC
after quenching reaction aliquot into H2O.
counterpart,6 is a significantly more active allylation catalyst
than B(C6F5)3 toward o-anisaldehyde (Table 1, entries 1 and
2).
To probe the origin of this unusual observation, the
thermoynamic data for the equilibrium involving ion pair
2b was acquired in a fashion analogous to that described
above for 2a (Scheme 2). As with the B(C6F5)3-promoted
reaction, all of the allylSnBu3 was rapidly and irreversibly
consumed in a C-C bond forming reaction with borane-
activated aldehyde. Equilibrium was reached after ∼5 h at
-80 °C. The only signals in the 119Sn NMR spectrum of the
equilibrium mixture are of stannyl ether 4 at 110 ppm and a
signal identical to that reported for the bis-aldehyde ligated
tributyltin cation [(o-anisaldehyde)2SnBu3]+ at 92 ppm.4 The
19F NMR spectrum indicates the presence of only two
fluorine-containing species; the borane-aldehyde adduct 1b
and a second set of signals due to the alkoxy borate species
2b.7 Finally, van’t Hoff analysis of this equilibrium (Figure
1, open triangles) yields a ∆S value of 18.3(1) eu and a ∆H
of 5.70(5) kcal mol-1. Not surprisingly, the ∆S values for
these two equilibria are quite similar, but the ∆∆H of 1.51-
(5) kcal mol-1 indicates that the ion pair 2b is less
enthalpically stable than the fully fluorinated alkoxyborate
2a. The least-squares analysis of the van’t Hoff plots allows
estimation of Keq values at one temperature for direct
comparison of the two boranes. At 193 K, Keq values are
estimated to be 3.0 × 10-3 for PhB(C6F5)2 and 6.0 × 10-5
for B(C6F5)3.
The ability to “turn on” path b via modification of Lewis
acidity has significant implications for development of this
chemistry, since chiral boranes can be expected to be
effective for asymmetric allylation procedures by this mech-
(5) Deck, P. A.; Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 1998,
120, 1772.
(6) (a) PhB(C6F5)2 exhibits a Child’s Lewis acidity6b (0.54) lower than
that of B(C6F5)3 (0.68). In addition to significantly weaker Child’s Lewis
acidity, in a competition experiment between B(C6F5)3 (1 equiv) and
PhB(C6F5)2 (1 equiv) for PhCHO (0.9 equiv) in CD2Cl2, only the B(C6F5)3-
PhCHO adduct 1a was observed by 19F NMR, confirming the higher
Lewis acidity of B(C6F5)3 for PhCHO (see Supporting Information).
(b) Childs, R. F.; Mulholland, D. L.; Nixon, A. Can. J. Chem. 1982, 60,
801.
(8) (a) At the same temperatures and times, use of separately prepared8b
stannylinium cation [Bu3Sn]+[B(C6F5)4]- as a catalyst gives conversions
of 40% at -44°C and 9% at -78°C. (b) Lambert, J. B.; Kuhlmann, B. J.
Chem. Soc., Chem. Commun. 1992, 931.
(9) (a) Unfortunately, the final member of this family of boranes, namely,
Ph2B(C6F5) is prone to redistribution,9b such that small amounts of PhB-
(C6F5)2 are always present, skewing comparisons between the two. (b)
Bradley, D. C.; Harding, I. S., Keefe, A. D.; Motevalli, M.; Zheng, D. H.
J. Chem. Soc., Dalton Trans. 1996, 3931.
(7) The spectroscopic signature of the alkoxyborate anion in 2b was
identical to that of its [NBu4]+ salt, generated separately (see Supporting
Information for details).
Org. Lett., Vol. 5, No. 16, 2003
2859

anism but not via the tin-catalyzed path a. We are pursuing
the synthesis of chiral perfluroaryl boranes and further
examining the effects of solvent and temperature on the
equilibria described herein to inform our design of new
LAs.
(Canada) and NSERC in the form of a CRD Grant, an E.
W. R. Steacie Memorial Fellowship (W.E.P. 2001-2003), and
a PGSA Fellowship (D.J.M.).
Supporting Information Available: Full experimental
and characterizational details and Keq data. This material is
available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. Funding for this work was provided
by the Merck Frosst Centre for Therapeutic Research
OL034928I
2860
Org. Lett., Vol. 5, No. 16, 2003